Natural daylight, as directly or indirectly provided by the Sun, changes in spectral composition over the day, due to changes in latitude and longitude of the Sun relative to an observer, which changes transmission and scattering paths in the earth's atmosphere, and reflection and scattering of objects near the observer. It is desired to recreate (at least to certain extent) these effects in artificial light sources, by changing the light sources' spectral composition and color of emission, or to be more specific, to change the correlated color temperature of its light output. Potential application would be in retail or residential environments, to change the lighting atmosphere as well as changing the mood and well-being of people. Additionally, it is desired to implement such functionality with only limited added cost, and minimum number of added components, while maintaining a high efficiency (luminous flux output compared to electrical power going in, while maintaining good CRI).
It is also desired to change the color point of solid state light sources which do not meet the target color point specifications. Such deviations for example occur due to production variations in wavelength or efficiency, or due to variations in phosphor conversion efficiency in case phosphors are used to create different spectral components of the light output. These conversion efficiencies can vary due to differences in layer thicknesses, or variations of the phosphor particle concentration in the phosphor layer (or layers), or due to variations in the chemical composition of the phosphor. In this case it is also desired to have the ability to adjust the color point of a solid state lighting module after it has been assembled, so that module meets color point targets.
It is known that modules can be made with strings of red, green, and blue light emitting diodes (LEDs), where each string is attached to a current source, and where each of the current sources can be adjusted to change the relative light output of the red, green and blue emitting LEDs, so that different shades of white or any other color can be produced. Some drawbacks of this approach are that multiple drivers are required, which increases the number of components needed and costs, and that only a portion of all the LEDs are used at full capacity at any given time. If, for example, light with a high correlated color temperature is desired, which has a relative high blue content, the blue LEDs are driven at maximum drive condition, while the green and in specific the red LEDs are driven at a current much below their typical drive currents. If however a light output with a low correlated color temperature is required, the red LEDs are driven to a maximum, while the blue LEDs are driven at a much lower current than typical. On average, the number of LEDs required is more than if the system would be optimized for only one color point.
Furthermore, due to varying drive conditions the efficiency of the LEDs varies (due to the so called current and temperature droop), which requires more electronics to predict the actual color of the light output in relation to the drive current. Typically this is done with a micro-controller, and very often additional measurements of for example the board temperature are required as inputs for the algorithms programmed in the micro-controller. This approach has an additional drawback, in that the devices suffer from differential aging. For example, red LEDs can degrade faster than the blue LEDs if they are driven harder, or blue LEDs can degrade faster, when the device is operated at relatively high color temperatures. With respect to differential aging the situation is even worse, since it is known that LEDs aging (degradation of the light output at same input power over time) can differ from device to device.
A solution for this is to use a technique where at least three sensors are used, each of the sensors having different spectral responses, and where the signals of the three sensors are measured and used to get an estimate of the actual color point of the output of the module. This measurement is then used to control the currents through the strings of red, green and blue LEDs using an electronic feedback control. Such a technique is commonly referred to as an optical feedback technique. Drawbacks of this approach include an increasing number of components, and the need of embedded micro-controllers, which of course results in additional costs, and increased chances of electronic failure.
Besides using red, green and blue light emitting diodes in these systems, combinations of other colors can be used, including white LEDs, or a combination of white LEDs having different correlated color temperatures.
An example of a system where white and red LEDs are used is the system produced by LED Lighting Fixtures (NC, USA), which was recently acquired by CREE (NC, USA). The system is a down-light module with a mixing cavity using yellow LEDs in combination with red LEDs to produce a warm white color, and a sensor which is used to measure the relative light output of the yellow versus the red LEDs, and to maintain a constant color for the light output of the down light. This system is not designed to change the color of light output at request of the user of the system, but the color can be set by adjusting the control conditions at the factory.